Vaccination: A Novel Approach to Reduce Antibiotic Resistance

Received 26 April 2004; accepted 27 April 2004; electronically published 16 August 2004. Reprints or correspondence: Dr. Keith P. Klugman, The Rollins School of Public Health, Emory University, 1518 Clifton Rd. NE, Rm. 720, Atlanta, GA 30322 (kklugma@sph.emory. edu). Clinical Infectious Diseases 2004; 39:649–51 2004 by the Infectious Diseases Society of America. All rights reserved. 1058-4838/2004/3905-0005$15.00 The state of Tennessee has had, for the past 2 decades, the greatest burden of antibiotic-resistant pneumococci isolated in the United States, culminating, in 1999, in rates of penicillin resistance as high as 60% among blood isolates from young children !2 years of age [1]. It is not just the proportion of penicillin-nonsusceptible strains isolated from blood that is remarkable, but also the degree of resistance that is due, at least in part, to the occurrence of a pneumococcal clone named for the state (Tennessee -4) [2]. The type strain of this clone has a cefotaxime MIC of 32 mg/mL—one of the most cephalosporin-resistant pneumococci ever described [2, 3]–and was originally isolated from the blood and CSF of a child in Memphis, Tennessee, whose cefotaxime therapy failed [4]. The situation in Tennessee is all now changing, perhaps partly because of improvements in empirical antimicrobial prescribing, although more likely, as argued by Talbot and colleagues [1], because of the introduction of pneumococcal conjugate vaccine (PCV). Since the introduction of the vaccine in 1999, the proportion of resistant strains has fallen dramatically, not only in immunized children, but also in nonimmunized older children and adults. Although the fall in the proportion of antibiotic-resistant strains among children (from 59.8% to 30.4%) is dramatic, it is not as dramatic as the reduction in invasive disease due to vaccine serotypes among children of the same age (from 106.3 to 13.8 cases per 100,000). Although the exact numbers cannot be calculated from the data presented, it is of concern that the proportion of resistant strains among residual conjugate-vaccine serotypes likely remains high, and selection of resistance in those and other serotypes is certainly continuing. The emergence of resistance in nonvaccine serotypes has been documented in the absence of conjugate vaccine selection in isolates from middleear specimens from Israeli toddlers [5], and previously rare vaccine serotypes, such as 35B, have also emerged as important antibiotic-resistant clones in the United States [6]. The data from Tennessee [1] provide some evidence for the low-level emergence of nonvaccine serotypes. Current models of the progression of antimicrobial resistance in the United States [7] suggest a continuous increase in antimicrobial resistance in the absence of a novel intervention, such as use of PCV, and the data from Tennessee clearly demonstrate the impact of such an intervention. Rapid reductions in resistance of this magnitude cannot be shown in resistance models to be due to changes in prescribing behavior alone [8]. How has the introduction of PCV had such a dramatic impact on antimicrobial resistance? The reasons are several. PCV interrupts the transmission of antibiotic-resistant strains by blocking the acquisition of vaccine serotype pneumococci that are resistant to antibiotics. The emergence of antimicrobial resistance in the pneumococcus has taken place almost exclusively among pneumococcal serotypes commonly carried by children, suggesting that selection is favored by the long periods of time that these serotypes are carried by young infants [9]. The first evidence that PCV was able to interrupt the acquisition of antibiotic-resistant strains came from a study of vaccination with 1 or 2 doses of a now discontinued—at the time, investigational—7-valent vaccine conjugated to Neisseria meningitidis outer membrane protein that was given to toddlers in Israel [10]. The incidence of antibiotic-resistant vaccine-type strains was reduced to 4% among the vaccinated children, compared with an incidence of 14%